RussellSpitzer commented on code in PR #1889: URL: https://github.com/apache/polaris/pull/1889#discussion_r2150344807
########## site/content/in-dev/unreleased/generic-table.md: ########## @@ -0,0 +1,168 @@ +--- +# +# Licensed to the Apache Software Foundation (ASF) under one +# or more contributor license agreements. See the NOTICE file +# distributed with this work for additional information +# regarding copyright ownership. The ASF licenses this file +# to you under the Apache License, Version 2.0 (the +# "License"); you may not use this file except in compliance +# with the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, +# software distributed under the License is distributed on an +# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY +# KIND, either express or implied. See the License for the +# specific language governing permissions and limitations +# under the License. +# +title: Generic Table (Beta) +type: docs +weight: 435 +--- + +The Generic Table in Apache Polaris provides basic management support for non-Iceberg tables. + +With the Generic Table API, you can: +- Create generic tables under a namespace +- Load a generic table +- Drop a generic table +- List all generic tables under a namespace + +**NOTE** The current generic table is in beta release, there can still be incomplete features and bugs. Please use it +with caution and report any issue if encountered. + +## What is a Generic Table? + +A generic table in Polaris is a structured entity that defines the necessary information. Each +generic table entity contains the following properties: + +- **name** (required): A unique identifier for the table within a namespace +- **format** (required): The format for the generic table, i.e. "delta", "csv" +- **base-location** (optional): Table base location in URI format. For example: s3://<my-bucket>/path/to/table + - The table base location is a location that includes all files for the table + - A table with multiple disjoint locations (i.e. containing files that are outside the configured base location) is not compliant with the current generic table support in Polaris. + - If no location is provided, clients or users are responsible for managing the location. +- **properties** (optional): Properties for the generic table passed on creation +- **doc** (optional): Comment or description for the table + +## Generic Table API Vs. Iceberg Table API + +Generic Table provides a different set of APIs to operate on the Generic Table entities while Iceberg APIs operates on +the Iceberg Table entities. + +| Operations | **Generic Table API** | **Iceberg Table API** | +|--------------|-------------------------------------------------------------|-------------------------------------------------------------| +| Create Table | create an Iceberg table | create a generic table | +| Load Table | load an Iceberg table, load a generic table throws an error | load a generic table, load an Iceberg table throws an error | +| Drop Table | drop an Iceberg table, drop a generic table throws an error | drop a generic table, drop an Iceberg table throws an error | +| List Table | list all Iceberg tables | list all generic tables | + +Note that generic table shares the same namespace with Iceberg tables, the table name has to be unique under the same namespace. + +## Working with Generic Table + +There are two ways to work with Polaris Generic Tables today: +1) Directly communicate with Polaris through REST API calls using curl. Details will be described in the later section. +2) Use the Spark Client provided if you are working with Spark. Please refer to [Polaris Spark Client]({{% ref "polaris-spark-client" %}}) for detailed instructions. + +### Create a Generic Table + +To create a generic table, you need to provide the corresponding fields as described in [What is a Generic Table](#what-is-a-generic-table). +Following is the REST API for creating a generic Table: + +```json +POST /polaris/v1/{prefix}/namespaces/{namespace}/generic-tables +{ + "name": "<table_name>", + "format": "<table_format>", + "base-location": "<table_base_location>", + "doc": "<comment or description for table>", + "properties": { + "<property-key>": "<property-value>" + } +} +``` + +Here is an example to create a generic table with name `delta_table` and format as `delta` under a namespace `delta_ns` +for catalog `delta_catalog`: + +```json +POST /polaris/v1/delta_catalog/namespaces/delta_ns/generic-tables +{ + "name": "delta_table", + "format": "delta", +} +``` + +### Load a Generic Table +The REST API for load a generic table is the following: + +```json +GET /polaris/v1/{prefix}/namespaces/{namespace}/generic-tables/{generic-table} +``` + +Here is an example to load the table `delta_table`: +```json +GET /polaris/v1/delta_catalog/namespaces/delta_ns/generic-tables/{generic-table} +``` +And the response looks like the following: +```json +{ + "table": { + "name": "delta_table", + "format": "delta", + "base-location": null, + "doc": null, + "properties": null + } +} +``` + +### List a Generic Table +Here is the REST API for listing the generic tables under a given namespace: +```json +GET /polaris/v1/{prefix}/namespaces/{namespace}/generic-tables/ +``` + +Following rest call lists all tables under namespace delta_namespace: +```json +GET /polaris/v1/delta_catalog/namespaces/delta_ns/generic-tables/ +``` +Example Response: +```json +{ + "identifiers": [ + { + "namespace": ["delta_ns"], + "name": "delta_table" + } + ], + "next-page-token": null +} +``` + +### Drop a Generic Table +The drop generic table REST API is the following: +```json +DELETE /polaris/v1/{prefix}/namespaces/{namespace}/generic-tables/{generic-table} +``` + +To drop the table `delat_table`, use the following: +```json +DELETE /polaris/v1/delta_catalog/namespaces//generic-tables/{generic-table} +``` + +### API Reference + +For the complete and up-to-date API specification, see the [generic-tables-api.yaml](https://github.com/apache/polaris/blob/main/spec/polaris-catalog-apis/generic-tables-api.yaml). + + +## Limitations + +The support for cross engine sharing of Generic Table is very limited: +1) Limited spec information. Currently, there is no spec for information like Schema, Partition etc. +2) No commit coordination or update capability provided at the catalog service level. Review Comment: A user is tying a table-identifier to a bucket of properties which an engine can use to resolve and read a table. For example, Let's say I want to make a Delta Lake table in Spark with Unity, or HMS, or whatnot. The Delta catalog plugin will essentially bundle up a set of properties that say Table X is at location Y and potentially has other properties that should be used when accessing this table (Split sizes, compute behaviors etc ...) . That is packed up and put in a Catalog be in HMS or Unity or whatnot. A user when using SQL can reference the Identifier and the Catalog Plugin within Spark will pull down the set properties and use them to initialize a Spark Format Specific plugin to use those properties. In Delta's case it will load up the Delta Lake plugin and use the properties in the Catalog as the Datasource Reader properties. The datasource will then load a Spark Table object and return it to the user (or use it in a query). In the case of Delta the only properties actually relating to the table's state are the "Location" and "format", everything else is essentially re-loaded from th e delta log found at the Location. If instead you had a Cassandra entity, it would have properties format="cassandra" and then a bunch of properties which would dictate connection parameters for the C* cluster. Actual information about the "state" of the table other than it's connection format can only be read by loading the appropriate plugin and using the properties to connect to the cluster. This is model used by Spark for interacting with all Spark sources that don't have an independent V2 Catalog implementation (and actually works for those with a V2 Catalog Implementation as well as long as they also have a V2 source register implementation). So the user benefit of Polaris providing this is essentially allowing Polaris to be used as a generic Spark catalog and gaining support for all DSv1 and most DSv2 sources without having an additional store for a Spark cluster. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: issues-unsubscr...@polaris.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
